From mboxrd@z Thu Jan 1 00:00:00 1970 From: Andi Kleen Subject: Re: [Bug 15836] Commit 6ad696d2cf535772dff659298ec7e7260e344595 breaks my SD card reader [197b:2381] Date: Tue, 4 May 2010 22:13:03 +0200 Message-ID: <20100504201302.GD13655@basil.fritz.box> References: <201004262048.o3QKml95025961@demeter.kernel.org> <20100426135800.28378e99.akpm@linux-foundation.org> <20100430170614.GA22807@gargoyle.fritz.box> <1272739661.10538.30.camel@weslok.olx.intra> <20100501230443.GB8074@gargoyle.fritz.box> <1272912441.7490.6.camel@weslok.olx.intra> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: Received: from one.firstfloor.org ([213.235.205.2]:46276 "EHLO one.firstfloor.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S933608Ab0EDUNE (ORCPT ); Tue, 4 May 2010 16:13:04 -0400 Content-Disposition: inline In-Reply-To: <1272912441.7490.6.camel@weslok.olx.intra> Sender: linux-mmc-owner@vger.kernel.org List-Id: linux-mmc@vger.kernel.org To: Stephan =?iso-8859-1?Q?S=FCrken?= Cc: Andi Kleen , Andrew Morton , linux-mmc@vger.kernel.org, bugzilla-daemon@bugzilla.kernel.org On Mon, May 03, 2010 at 08:47:21PM +0200, Stephan S=FCrken wrote: > On Sun, 2010-05-02 at 01:04 +0200, Andi Kleen wrote: > (..)=20 > > > config-2.6.33.2.git19f00f0-config-2.6.33-2-amd64.nohiber:CONFIG_M= EMORY_HOTPLUG=3Dy > > > =3D>"bad" > > > config-2.6.33.2.git19f00f0-config-2.6.33-2-amd64.nohotplug:# CONF= IG_MEMORY_HOTPLUG is not set > > > =3D>"good" > (..) > > Can send the diff of two boot logs with memory hotplug on and off? > > I wonder if you have the same memory layout in both cases. >=20 > Attached, from the above two good/bad kernel builds; I stripped the > stamps to make it more readable. Thanks, I was hoping for a comparison between memory hotplug on and off with hibernation always off, this seems to be between memory hotplug and=20 hibernation.=20 Anyways looking at this log. The only significant difference I can see is that in the one case a lot of stuff including sdhci=20 get initialized earlier. Presumably the timing is somewhat=20 different related to the USB scanning and the memory hotadd kernel being a bit slower to boot (perhaps due to the increased printks= ) Perhaps this exposes some races that cause the problem. What these races are I don't know. I would suggest to check with the mmc maintainer. -Andi